52 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 70 



V. THE VARIATIONS IN THE SURFACE TEMPERATURE 



As already noted, our curves (figs. i6 to 19) for the temperature 

 anomalies of the ocean surface in the single 2° longitude fields in 

 the whole eastern part of the investigated region show a marked 

 agreement over great regions. The dififerent characteristic lines, 

 which indicate the variations, change from field to field by a gradual 

 march as o'ne looks forward, for example, from the east toward the 

 west. Among prominent common features may be noted the depres- 

 sion in the year 1904, which appears on all the eastern curves for 

 the first decade group February 3 to March 4, particularly from 

 10° west longitude to 40° west longitude, and in part even to 50° 

 west longitude. An equal depression is found in the curve for the 

 second decade group (March 15 to April 13), in the eastern part 

 of the investigated region. But further toward the west the great- 

 est depression occurs in the year 1903. 



In common for a considerable part of the curves as well for the 

 first as for the second decade groups, is also a depression occurring 

 in the year 1899. The curves tend toward a maximum in the year 

 1901. Further on the temperature generally rises strongly from 

 the year 1904 with small breaks to the years 1906, 1907 and 1908. 



The curves for the single 2° fields westward of 44° and 46° west 

 longitude at 41° north latitude show apparently only slight cor- 

 respondence, and our figures 17 to 19 are calculated to give an 

 impression of a chaotic medley of lines with no similarity. That 

 these irregularities begin at about 46° west depends upon the fact 

 that here a certain great discontinuity in temperature of the fields 

 occurs from temperatures about 13° or 14° C. to between 6.8° and 

 8.5° C. The irregularities in the curves westward of this boundary 

 have clearly as their cause the fact that in this part of the ocean 

 the isotherms for the ocean temperatures lie so closely together 

 that a comparatively slight difiference of locality even within the 

 same 2° fields is sufficient to produce a great temperature change, 

 so that the distribution of the observations within the field may 

 have a great influence on the mean value. Besides this, it occurs 

 that inaccuracies in the determination of the position of the observer 

 may produce a noticeable influence in this region. Furthermore, 

 slight local movements of the water surface may easily produce 

 changes in the surface temperature in such localities. 



It is moreover probable that many accidental errors may play a 

 part in the computed mean values for the single fields, even when 

 the number of observations is quite large. More than a moderately 



